Author + information
- Received December 5, 2002
- Revision received January 30, 2003
- Accepted February 20, 2003
- Published online October 1, 2003.
- Joseph L Blackshear, MD*,* (, )
- W.Dudley Johnson, MD†,
- John A Odell, MD*,
- Vickie S Baker, RN*,
- Mary Howard, RN†,
- Lesly Pearce, MS‡,
- Christopher Stone, MD†,
- Douglas L Packer, MD§ and
- Hartzell V Schaff, MD§
- ↵*Reprint requests and correspondence:
Dr. Joseph L. Blackshear, Mayo Clinic Jacksonville, 4500 San Pablo Road, Jacksonville, Florida 32224, USA.
Objectives We evaluated left atrial appendage obliteration in high-risk patients with atrial fibrillation (AF).
Background Left atrial appendage thrombosis and embolization is the principal mechanism of stroke in AF. Anticoagulation is underutilized and often contraindicated.
Methods Thoracoscopic Left Appendage, Total Obliteration, No cardiac Invasion (LAPTONI) was undertaken with a loop snare in eight patients and a stapler in seven patients, median age 71 years, with clinical risk factors for stroke and with an absolute contraindication to or failure of prior thrombosis prevention with warfarin. Eleven patients had a history of prior thromboembolism. One patient took sustained warfarin during follow-up.
Results The LAPTONI procedure was completed in 14 of 15 patients, and 1 patient required urgent conversion to open thoracotomy because of bleeding. Patients have been followed up for 8 to 60 months, mean 42 ± 14 months. One fatal stroke occurred 55 months after surgery, and one non-disabling stroke three months after surgery. Two other deaths occurred, one after coronary bypass surgery and the other from hepatic failure. The subgroup of 11 patients with prior thromboembolism had an annualized rate of stroke of 5.2% per year (95% confidence interval [CI] 1.3 to 21) after LAPTONI, which compares to a rate of 13% per year (95% CI 9.0 to 19) for similar aspirin-treated patients from the Stroke Prevention in Atrial Fibrillation trials (p = 0.15).
Conclusions The LAPTONI procedure appears technically feasible without immediate disabling neurologic morbidity or mortality, and it demonstrates low post-operative event rates and a statistical trend toward thromboembolic risk reduction in high-risk AF patients.
The stroke rate in patients with atrial fibrillation (AF) older than 75 years of age is 8.1% per year with one clinical risk factor and is 12% per year at any age in clinical trial populations with a history of prior thromboembolism (1,2). Three-year stroke rates in elderly nursing home patients not anticoagulated are in excess of 50% (3). Fear of hemorrhagic complications prevents utilization of anticoagulation in these high-risk patients.
Most thrombi in non-rheumatic AF originate in the left atrial (LA) appendage (4), suggesting the origin of thromboembolism (5). Left atrial appendage obliteration is commonly performed in a variety of cardiac surgical operations, but its feasibility and value as a stand-alone stroke preventative strategy is undefined. A catheter-based LA appendage occluder device has recently been developed and tested in animals, and a first group of humans has undergone device implantation (6). However, there are no data to support the hypothesis that appendage obliteration reduces stroke. We present data on isolated surgical obliteration of the LA appendage in 15 patients with a mean follow-up in excess of three years. Data on seven of these patients were previously published shortly after surgery (7).
Patients at two medical centers with chronic or intermittent AF were identified from routine clinical practice with at least one risk factor for stroke (prior thromboembolism, hypertension, age >75 years, left ventricular dysfunction, or diabetes mellitus) andeither an absolute contraindication to warfarin or documentation of prior LA appendage thrombosis despite adequate oral anticoagulation. The protocol was approved by the institutional review boards of both institutions.
Thoracoscopic closure of the LA appendage
Informed consent was obtained before surgery. Each patient was anesthetized and intubated with a double lumen endotracheal tube. Transesophageal echocardiography (TEE) confirmed the absence of LA appendage thrombus at the time of surgery. Patients were positioned, prepared, and draped as for a left or left-lateral thorocotomy. Three to five thoracoscopic incisions were made to introduce the video-assisted thoracoscopic instruments. Using a double-lumen endotracheal tube, the volume of the left lung was reduced. Up to 4 cm pressure CO2insufflation in the left pleural space was used specifically to reduce or eliminate the potential for air embolus in the event of an inadvertent tear. The pericardium was opened with scissors at a site parallel to the phrenic nerve. The atrial appendage tip was grasped using a non-traumatic clamp, and an endoloop snare was fixed as a tie to the appendiceal tip. A second endoloop was introduced over the first tie and manipulated with the assistance of instruments to the appendiceal base, where it was cinched to occlude the appendage. Alternatively, a row of staples was placed. Occlusion was documented by TEE as the abrupt appearance of dense, stationary, high-contrast echoes (Fig. 1). At the conclusion of the procedure, the pericardium was left open, the CO2was evacuated, and tube thoracoscopy suction was employed.
Patients were regularly contacted after the procedure and questioned regarding neurologic symptoms. Contact was by telephone interview at six-month intervals. Medical records from hospitalizations were reviewed. The protocols did not include late imaging for verification of persistent occlusion of the LA appendage.
Patient exposure for stroke was computed as the number of days from LAPTONI to date of first stroke, with exposure censored on the date of death or last day of follow-up if no stroke occurred. Similarly, patient exposure for death was computed as the number of days from LAPTONI to the date of death, with censorship on the last day of follow-up. Overall event rates were calculated by dividing the number of events by the total patient observation, with 95% confidence intervals calculated using a Poisson distribution. Survival curves were computed using the Kaplan-Meier method and compared to reference populations. All statistical tests were two-sided, and statistical significance was accepted at the 0.05 level.
Patients' characteristics are listed in Table 1. Ages varied from 24 to 81 years, mean ± SD 67 ± 15 years, median 71 years. Eleven patients had a history of prior thromboembolism, 12 had a history of hypertension, and 11 had bleeding diatheses including five with prior intracranial hemorrhage and one with prior meningioma surgery. Hospital length of stay varied with from two to 20 days. Post-operative complications included bleeding from a torn accessory atrial appendage, necessitating urgent thoracotomy in one patient. One patient with hypertrophic nonobstructive cardiomyopathy experienced refractory pump failure and a prolonged intensive care unit stay after the procedure. One patient had atelactasis, one had prolonged air leak, one underwent investigation and treatment of genitourinary bleeding, and one had chronic pleuritic pain. There were no peri-operative neurologic events.
Median follow-up duration was 42 months. During this time, two strokes—one non-disabling and one fatal—occurred. One patient with a history of multiple prior strokes had a small stroke during hospitalization for pneumonia. He made a complete recovery and remains well 40 months after surgery. One patient, age 79 at surgery, with prior stroke, congestive heart failure, severe left ventricular dysfunction, and multiple subdural hematomas, suffered a fatal stroke and intracranial hemorrhage 55 months after surgery. One patient died after coronary bypass surgery, performed 34 months after LAPTONI. One patient with chronic hepatitis C, variceal and small intestinal blood loss, and prior systemic embolism died from hepatic failure. Overall rate of stroke was 4.0% per patient-year, 95% confidence interval (CI) 1.0 to 16. Rate of fatal stroke was 2.0% per patient-year, 95% CI 0.3 to 13. Rate of death or stroke was 6.0% per patient-year, 95% CI 2.0 to 19. Among the 11 patients with a history of prior thromboembolism, stroke rate was 5.2% per year, 95% CI 1.3 to 21, which compares to 13% per year, 95% CI 9.0 to 19 in aspirin-treated patients with prior thromboembolism in the Stroke Prevention in Atrial Fibrillation (SPAF) studies (p = 0.15, not significant) (8). In four patients without a history of prior stroke, the stroke rate was 0% per year after LAPTONI, compared to a rate of 2.6% per year, 95% CI 1.9 to 3.6% per year rate of stroke in SPAF (not statistically significant). Figure 2presents cumulative freedom from thromboembolism with comparison to rates in prior studies of similar patients (1,2).
The recent publication of a catheter-based LA appendage occluder device and its use in 30 patients (6)and the absence of data supporting the safety and efficacy of such a procedure provided us the impetus to present our results. Our two-center experience utilized surgical appendage obliteration with either a loop snare or a stapler, with placement of the occlusion device at the appendiceal base. Pre-clinical experience with both techniques documented a smooth LA architecture with complete obliteration (9). The site of closure is similar to that documented with the transvascular occluder (6)—but without the foreign body in immediate contact with the circulation. Although underpowered for statistical significance, the observed event rate in our series of patients with a history of prior thormboembolism is consistent with the benefits proposed for appendage obliteration. This observation would appear to support the further development and testing of strategies to reduce stroke risk in this highest risk segment of the AF population (1,2)in whom the preferred therapy, anticoagulation, is contraindicated or has previously been insufficient to prevent embolism. Clearly the magnitude of risk varies but may be very high. In a study of secondary stroke prevention in AF, non-anticoagulated patients had a 17% per year rate of stroke or death (2). Age, itself a significant AF stroke risk factor, was lower in the anticoagulation clinical trials than the AF population median of 75 years, suggesting that clinical trial data underestimate the true risk for this group. Consistent with this, Aronow et al. (3)described 312 persons with AF with an average age of 84 years who resided in a chronic care facility. Rates of stroke over three years in those not anticoagulated were 56% in those with no prior history of thromboembolism and 81% in those with prior history of thromboembolism (3).
Definitive proof that appendage obliteration by any technique reduces the risk of thromboembolism can be obtained only through a randomized clinical trial. However, data from Cox et al. (10)support the hypothesis that appendage ligation, a part of the maze procedure, may be effective in preventing stroke. More than 40% of patients who undergo the maze operation have transient post-operative AF, yet only those with mitral valve replacement, mechanical aortic valve replacement, or Benthall procedure are routinely anticoagulated. Despite this high rate of post-operative AF, the rate of post-operative stroke in the maze procedure population was 0.7% (10). Rates of peri-operative stroke for coronary bypass associated AF without maze vary from 2% to 7% (11), and post-operative AF was a highly significant independent predictor of stroke in the study of Stamou et al. (12). In addition, despite some reduction in atrial function in a portion of the post-maze population, during follow-up of between three months and 11.5 years, the stroke rate has been 0.4% (10). Patients undergoing maze surgery were relatively young and probably had a lower intrinsic risk of thromboembolism.
The obvious limitation of this study is the small numbers of study patients, yielding very wide CIs for event rates and preventing a definitive statement regarding efficacy. While they suggest a trend toward benefit, the event rates after LAPTONI fall within the confidence limits reported in clinical trials with aspirin or no therapy. Second, the absence of a simultaneous randomized control group prevents a truly accurate assessment of risk reduction. Despite the absence of mortality or neurologic morbidity from LAPTONI in the short term, we did experience significant, although not unexpected, non-stroke morbidity in our small series, suggesting that refinements in the procedure or more careful patient selection may be required. Cardiac failure in our patient with hypertrophic cardiomyopathy, and the prior report from Cox et al. that two of 10 peri-operative deaths in patients undergoing the maze procedure occurred in patients with hypertrophic cardiomyopathy (10), suggest that these patients should be excluded from the initial application of this procedure.
Surgery to prevent stroke is widely accepted, as in the case of carotid endarterectomy. Our initial human experience with LAPTONI in AF documents that it is technically feasible to perform this procedure without cardiac invasion, in a relatively elderly population at substantial risk for AF-related stroke, without producing any disabling neurologic morbidity in the short term. Our study differs from the endovascular approach thus far described (6), in that our patients had higher estimated risk, that no cardiac invasion or intracardiac foreign body was placed, and that no immediate post-procedure antithrombotic therapy was utilized. A statistical trend toward benefit in our patients with prior thromboembolism was demonstrated with an average follow-up of 42 months. These data provide support for the further development of this or other techniques for appendage obliteration, in hopes of providing an effective prophylaxis for patients with the highest risk and absolute contraindications to, or proven inefficacy of, anticoagulation.
☆ Supported by grants from Ethicon Endosurgery and the Mayo Foundation for Medical Research.
- atrial fibrillation
- Atrial Fibrillation Investigators
- European Atrial Fibrillation Trial
- left atrial
- Left Appendage, Total Obliteration, No cardiac Invasion
- Stroke Prevention in Atrial Fibrillation
- transesophageal echocardiography
- Received December 5, 2002.
- Revision received January 30, 2003.
- Accepted February 20, 2003.
- American College of Cardiology Foundation
- Hart R.G.,
- Halperin J.L.
- Sievert H.,
- Lesh M.D.,
- Trepels T.,
- et al.
- Johnson W.D.,
- Ganjoo A.K.,
- Stone C.D.,
- et al.
- Hart R.G.,
- Pearce L.A.,
- McBride R.,
- et al.
- Blackshear J.L.
- Stamou S.C.,
- Hill P.C.,
- Dangas G.,
- et al.